Wildfire threatening the power grid that supplies San Francisco | Hacker News Reader

Wildfire threatening the power grid that supplies San Francisco

npr.org

141 pointsmpchlets61 comments

For anyone wondering what the relationship between wildfires in Yosemite and San Francisco proper, its worth a small comment. California's Sierra Nevada range is a couple hours east of SF. These mountains include the highest points in the continental US. Trapping most of the water from the east-west jet stream. The sierras hold massive amounts of high-quality water, which runs out of the non-porous terrain of the high-alpine regions. The sierras are something of a geological monolith, not unlike a giant bathub [1]. SF taps into this water source at a dam on the West; LA taps into it on the East (LA acqueduct). This fire is threatening the a region about an hour or so north of Yosemite Valley, and west of Tuolome Meadows, where some critical infrastucure resides.[2]

[1] https://upload.wikimedia.org/wikipedia/en/a/a2/Sierra_nevada...

[2] https://en.wikipedia.org/wiki/O%27Shaughnessy_Dam_%28Califor...

Yikes! From the article:

"The Associated Press says San Francisco gets 85 percent of its water from the Yosemite-area Hetch Hetchy reservoir that is about 4 miles from the fire."

The fire is well established now, and in addition to the threat to the water supply they've already had to shut down electric transmission lines.

I remember talking to one of the rangers at Yosemite, near one of the natural lakes that eventually supply water to SF. I remember him saying that he tells visitors who go answer nature's call at the lake to think twice, since they're going to be drinking this very water soon.

What did he say about drinking fish shit? And the effluent of every other critter in the park?

I think that it has more to do with the idea that humans moving through that area would do much more damage (than the local animal population) if they were not discouraged in some way.

Rangers should discourage people from fouling the area, but the comparison made by the Ranger is nonsensical on a number of levels. Anyone who heard it might disregard it after recognizing how silly an argument it was.

Just as astronauts on the ISS drink their own piss, it is nothing but in our minds.

Drinking your own piss is different than drinking someone else's piss. :P

Urination is a relatively low-impact activity on a water supply. Defecation, on the other hand, tends to cause a decent variety of diarrheal illnesses unless the water treatment facility is quite good - something that may not be accurate for something designed for sterile snow runoff. The water here is not just being drunk, but going on fresh produce as well, so you can't rely on purely municipal filtering.

The SF water supply has some interesting politics:

https://www.baycitizen.org/news/water/raise-rent-hetch-hetch...

Water is intensely political pretty much everywhere but rainforests, and at all levels of government. However, unless you're directly involved, you don't hear about it because it can't normally be used as a political football. Your opinion on any particular water dispute will have nothing to do with your political views or party affiliation, it's mostly a product of geographic location or industry, and no politician who happens to represent both parties wants to take sides. (And any politician who represents only one side will invariably adopt that side's position.)

(Incidentally, in the modern era, the US Supreme Court's original jurisdiction case load is made up almost entirely of states suing one another over either water rights or boundary disputes based on rivers.)

We're not quite Mad Max yet. Perhaps 80% of the country lives in an area where water is not scarce enough to prompt water table depletion or aqueduct building. "Water politics" in these areas mostly consist of bureaucratic squabbles regarding the minor expense of water treatment facilities & pipe-building, or soft-science environmental debates about erosion, fish, and effluents - things which can be reliably consigned to the higher end of Maslow's Pyramid.

About half of the remainder of the problem could be solved easily by ending this stupid obsession for English Country Manor lawns well outside a climate zone where they're viable, and prohibitions on all but drip irrigation for agriculture. These could be mostly accomplished without legislatorial nitpicking by simply making prices reflect scarcity, permanently. We seem to have an innate resistance, politically, to pricing water to reflect its infrastructural and depletion-replacement cost, to admitting that some mechanism needs to scale back use.

Some issues are simply not solveable, it's true - isolated water tables that have been changed by taprooted invasives are probably never going to revert to grassland, and need total cessation of irrigation activities to avoid desertification. These cases are relatively rare though.

I've read that when California's major water systems were built long ago there were explicit compacts between the urban and agricultural regions. The latter now see various actions as breaking that.

This article is vague (and overly alarmist); other sources I have read indicate that the threat is to SF's municipal power company, which provides power primarily to municipal buildings (including, most notably, SFO!). The vast majority of people in SF and the Bay Area get their power from PG&E, which has power plants in many more places than just Yosemite.

More info: http://www.sfwater.org/index.aspx?page=391

Yes, the NPR article is alarmist, verging on irresponsible.

A few facts, with sources:

- The water coming out of Hetchy Hetchy is still well within the 'usable' range [1]

- The total loss of power is 293 MW (Holm and Kirkwood), which is a tiny percentage of the power generated in CA [1,2]

- The SF PUC is already dealing with the loss of those 293 MW: San Francisco is making up the difference in power generation by accessing power in an existing power bank and purchasing power on the open market. [1]

- PG&E is reporting no problems in their network (at 10:20 PT) [3]

Keep calm, carry on. Perhaps take it as an opportunity to turn of the lights you're not using, like you should every day ;)

1. http://www.sfwater.org/index.aspx?page=711 (see chart) 2. http://en.wikipedia.org/wiki/List_of_power_stations_in_Calif... 3. http://www.pge.com/en/myhome/customerservice/energystatus/gr...

Its only alarmist if it doesn't get worse :-) Seriously though, alarmist or not, its serious. This fire is growing and not contained. There exists a non-zero chance of it creating a serious impact into the available clean water supply. And this is to a county not exactly known for its fiscal responsibility or ability to respond well to crisis.

But at least San Francisco has maintained it's charm! That's what's really important.

To that point nothing about it on either news.google.com or the front of the NYT or WSJ (but there is something about graphene being a wonder material that is igniting a patent frenzy.)

Hetch Hetchy water doesn't just go to San Francisco. Two-thirds of it goes to other cities and towns on the Peninsula, the South Bay, and in Alameda County. In particular, many cities that the Hetch Hetchy Aqueduct runs through get some of the water, because voter approval in those cities was part of the campaign to get the project built.

For example, 74% of Mountain View's water comes from Hetch Hetchy. (They get 87% of their water from the SF Public Utilities Commission, and 85% of that comes from Hetch Hetchy.) [1]

Daly City, San Bruno, and South San Francisco currently get 67% of their water from SFPUC and the rest from local aquifers, with a project in the works to increase the SFPUC portion to 100%. [2]

Palo Alto, Menlo Park, and other towns also rely on it, although I don't know the percentages.

Many have seen the pipeline that crosses the Bay between the Dumbarton Bridge and the old railroad bridge.

There are also a number places in the Peninsula and South Bay where you can see parts of the Hetch Hetchy Aqueduct, either the pipes themselves or the above-ground access hatches where the pipe is underground. These are white structures, often circular.

For example if you walk or run the Dish Trail at Stanford, you can see a number of these next to the northernmost part of the trail. This is part of the southern branch of the Aqueduct, which doesn't cross the Bay but cuts south through Milpitas, Sunnyvale, and Mountain View.

Several places where you see a residential street that is divided with a very wide grassy or dirt median, that's the Aqueduct: Sharon Heights Drive and Ivy Drive in Menlo Park are examples.

Along Edgewood Road near 280 there are a number of pipeline sections where it alternates between above-ground bridges and underground sections through those hills. The pipeline then parallels the Cordilleras Trail in the Pulgas Ridge Open Space Preserve where you can see some of the access structures.

On the other side of 280 on Cañada Road there is the famous Pulgas Water Temple [4], where you can really get a sense of how much water flows through the Aqueduct.

For the obsessively curious like me, several years ago I traced the path of the southern branch of the Aqueduct and made a KML file marking some of the visible structures. [3] Someone else had made a similar file for the northern branch but I don't know where to find that now.

[1]: http://www.mountainview.gov/city_hall/public_works/water_con...

[2]: http://www.sfexaminer.com/sanfrancisco/three-peninsula-citie...

[3]: http://mg.to/earth/hetchy.kml

[4]: http://en.wikipedia.org/wiki/Pulgas_Water_Temple

I forgot to add... Here's a Street View from Ivy Drive in Menlo Park that happened to be taken during a pipeline addition/reconstruction. They set out the pipe sections in the street median before digging to lay them in place. These are some big pipes!

https://www.google.com/maps?ll=37.47579%2C-122.1629239999999...

(You may need to zoom out; I couldn't figure out how to save a URL with the correct zoom level.)

Looks like a good case for decentralization of the power grid; a good application for a small modular LFTR[1] (or similar) power units. As a bonus, you won't need to spoil the scenery with HV transmission lines anymore.

[1] Liquid Fluoride Thorium Reactor

Interesting concept. Of course, the first thing that comes to mind is how to protect these units from the bad guys.

If you mean "protect these units from bad guys so they don't threaten our electrical system," that's kind of the point of having a decentralized system of smaller reactors: if one or two get sabotaged, others in the system can easily pick up the slack.

If you mean "protect these units from bad guys so they don't steal the thorium and weaponize it," that's one of the great things about thorium reactors: thorium and its byproducts are very hard to weaponize. [1]

[1] https://en.wikipedia.org/wiki/Liquid_fluoride_thorium_reacto...

To which I say BS, at least for the reason cited.

Sure, the U-232 contaminated U-233 is nasty, but as long as it doesn't make it impractical to make a nuclear warhead it's quantitatively different from the uranium cycle, where after a few months at most plutonium is impossibly contaminated with two even more undesirable isotopes (one is very hot, I've seen estimates of 100kW for a bomb sized quantity (it's used for RTGs in deep space probes), the other precludes much of a bang and required the Manhattan Project to go with an implosion design).

Despite the gamma ray emission drawbacks, it could still be the easiest way to get lots of weapons grade fissionables from civilian power plants.

These folks seem to think the problem is solvable by adding U-238 http://www.coal2nuclear.com/MSR%20-%20Denatured%20-%20CNSLeB...

Yeah, that solves it neatly.

Reduces to a protoactinium problem, it decays to U-233 with a half-life of 27 days. The article claims there's so little protoactinium in the total mass of salt and stuff that it's not practical to isolate it, at least not without detection, and failing that, not quickly. This gets into fine details beyond my level of expertise, but I agree the problem is much reduced. Although very possibly still greater than for current LEU designs.

What about the health impact following a release of materials?

Molten salt thorium reactors are "continuous reprocessing" ones, which means they burn all the (active) waste they create for more energy.

Solid thorium reactors mix Th with the existing waste piles of Plutonium sitting in cooling pools and burn that up for us[1].

There's no waste materials to release from either of these reactors.

[1] http://www.extremetech.com/extreme/160131-thorium-nuclear-re...

Waste != Materials.

These reactors by definition have a fair amount of very "hot" materials; it hardly matters if none of them are declared "waste" if they'll still kill you in a few minutes of direct exposure.

...

molten salt reactors (the kind in the comment at the top of this thread) are at atmospheric pressure, and are designed with a drain plug which isolates the (very hot) materials from the neutron source, after which they cool down happily on their own in a separate but similarly shielded compartment.

release of materials is one of the most unlikely outcomes ever.

You really can't imagine a release of materials outcome in the context of active sabotage?

(Which is relevant in the context of massive distributed ones; current nuclear power systems mitigate this by being few in number such that they can be well guarded.)

since you can't weaponise the mixture, and it's just stupidly hot and dangerous for a while and then inert, I can't see why anyone would try to sabotage one. (apart from terrorists just out to cause mayhem, but that argument also works for cars, roads, planes, train stations, etc right?)

The addition of uranium or plutonium is unnecessary for operation of a LFTR, rather it is being sold as an extra advantage that a LFTR can safely consume the waste already created at PWR's. Consuming PWR waste in a LFTR is a better option than burying it in the desert (at least it is to some).

How would the materials be released? The reactors operate near 1 Atmosphere. If they are switched off, the molten salt cools and solidifies.

There could be a risk of contamination from the cooling loop, but such contamination might even be less than the normal radioactive fallout from the flues of coal plants.

http://en.wikipedia.org/wiki/Shaped_charge

I've been trying to imagine how blowing one of these things up with shaped charges, or anything else, is worse than blowing up other things like industrial plants, gas stations, etc.

The radioactive fuel is a molten salt, so it will cool and solidify soon after dispersal. It is certainly not good, but is it really worse than say, a chemical plant?

The radioactive stuff won't remain airborne after the initial blast, so that means a predictable and small area to remediate.

The fuel dispersed in an explosion can be collected, since it will solidify, so it won't poison the water table.

While not perfectly safe (nothing is), the disaster contingencies seem fundamentally different and better than those from a PWR + solid nuke waste disposal site.

There would be many more of them, which is both good and bad for various cases.

That's an answer to how almost anything can be released.

Petrol from a tanker? Shaped charge.

Cracking armored vehicle? Shaped charge.

Unlocking stuck door? Shaped charge.

It really does seem people try to downplay any potential reactor designs because global thermonuclear war may damage the reactor. Stop living in fear.

There are two issues here, one of which strongly cuts against distributing nuclear power that broadly:

If you have a lot of units, the loss of one or a few has limited effects, vs. taking out a few big baseline power plants and/or major transmission nodes.

What's the local/regional cost if someone takes out a node with a sufficiently big shaped charge?

If inter-connectivity is maintained at local power system boundaries, then adjacent nodes can pick up the slack. Another good argument for a power grid of peers is that nodes can be taken offline for maintenance without causing a disruption.

I'd hope the first thing to come to mind is how to build the technology. Once it's built, we can start worrying about how to protect it .. why protect something that doesn't exist?

Why build something if you can determine you can't adequately protect it?

Why build 3d printers if they could be used to make guns? Why invent bitcoin if it could be used to fund terrorism?

You make things because they are interesting to make, and could have a positive impact in the world. Protecting yourself from "the bad guys" is a secondary (or even tertiary) concern, unless the very purpose of the thing is to protect people from bad guys (for example a military fort).

Why live when you could die suddenly?

This is the kind of ethos that prods people to bolt on security rather than bake it into their code.

The system continues to be broken. Hundreds of millions going into building lethal drones to deploy to Afganistan/ME; some major american cities are on a brink of releasing police surveillance drones en masse (its constitution/law-questionable now), and yet we do not have drones to fight fire.

I am pretty sure there are lot of challenges within the project, but I can envision drones working together to connect each other and transport water to sprinkle it when needed without humans being endangered. I mean, if we achieved this [1], why not go further?

[1] http://www.youtube.com/watch?v=oDyfGM35ekc

Errrm, this translates into a "making large drones" problem. Which has a whole set of additional problems, e.g. what are you going to say to the survivors of a devastated neighborhood when a big drone with lots of water crashes into it?

I'm not saying its insurmountable, just that it's way beyond the current levels of both drone technology and acceptance.

The biggest problem fighting fires are lethal to humans high temperatures and lack of easy access to places affected. Those two problems would be irrelevant to properly designed drones that can easy maneuver and are built from temperature-resistant materials.

Instead of huge drone carrying water (we already have planes to do that now), I was envision a network of connected drones that "pass by" the water as they hover in the line all the way from water pool to fireplace. The video attached previously shows you group of drones perfectly communicating with each other.

I never said it was easy to develop, but I have to disagree its something "way beyound the current levels" of technology. Further, the cost to design, develop and implement would be pennies comparing to an average damage of mid-size fire. Not the mention about life-loss.

You either need large payloads or large numbers, but Thousands of drones flying around would likely lead to accidents. Accidents woudl likely lead to new fires, considering most of these things are happening in areas that are tinderboxes with significant aggrivating factors (high winds, exposure, etc).

The strategy for containing these fires is typically fuel deprivation, rather than "flame extinguishment" which is so difficult if not impossible for the above reasons. The guys that jump out of planes to fight backcountry fires, for example, are armed with axes and saws. Absent something like a star-wars (AT-AT) its not clear a direct frontal confrontation is pyhsically feasible, even from the air, for many fires.

Isn't there a Google interview question for this?

Damn. The powers that be better do something about this fast. SF is about one of the last places in the US to develop new things.

Heck just a temporary prolonged outage of Google alone would have huge negative consequences for the US (not to mention the world). Amazon AWS US west is in Northeren California. I wonder how big the negative effect will be on just how many companies?

The "City" of San Francisco (though it does contain a big chunk of the tech industry and houses many of it's employees) is not the whole "Silicon Valley".

Google, Facebook, LinkedIn, Y-Combinator, Stanford, Yahoo, etc are all located in the South Bay and (may or may not be, the article doesn't specify) affected by the fires.

Don't know about the power grid, but about half the water supply in Mountain View comes from local wells, so it would not be as affected as SF, but part of the rest comes from Hetch Hetchy, so there could still be problems. Not sure if local resources get reallocated in emergencies.

http://www.valleywater.org/Services/CleanReliableWater.aspx

Most people think hackers don't have good personal hygiene skills anyway, so I think those companies could skate for a little bit if push came to shove...

Most of the 'Silicon Valley' cities get at least some of their water from the Hetch Hetchy system.

This report is an economic scenario analysis, and estimates the probable effects on the Bay Area economy of a major failure of the San Francisco Public Utility Commission’s Hetch Hetchy water system,

http://www.bayeconfor.org/pdf/hetchhetchyfinal2.pdf (map on pg 4)

> SF is about one of the last places in the US to develop new things.

Don't believe your own hype. It's bad for you.

This sort of thing is EXACTLY WHY both Google and Amazon have datacenters all over the world.

Sure, except us west is only in NC and many businesses aren't likely to shift availability zones.

If you host a site that needs five nines or whatever in only one availability zone, it's your own darn fault.

Is there a sim where I can go an blow these places up in a simulator, eg anti terrorist training ?

If you mean "cyber-blow-up" there's https://www.sans.org/netwars/cybercity

I've no idea if it lives up to its own hype.

anti terrorist training? I suspect these simulators aren't more commonplace precisely because they may be accused of being training software for terrorists.

Reply on news.ycombinator.com